Well, what we should really do is document exactly whats going on, so we can then write tests to ensure the code matches the intent.
Adrian 2008/7/17 Mark Nottingham <[EMAIL PROTECTED]>: > Hmm, are you saying that it may have behaved differently in the past? If so, > we should probably treat it as a regression; if not, a new config option. > > On 16/07/2008, at 10:37 AM, Adrian Chadd wrote: > >> Sure. >> >> It doesn't surprise me, a lot of the acceleration/vhost stuff was >> shuffled around between 2.6 and 2.7; I doubt any of it is documented >> either making testing quite difficult.. >> >> (Squid-core guys, can we please sort out whats needed to bring Mark on >> as a committer already?) >> >> >> Adrian >> >> 2008/7/17 Mark Nottingham <[EMAIL PROTECTED]>: >>> >>> From what I can see on the wire, Squid 2.7 does not rewrite the Host >>> header >>> when sending a request to an origin server when running in vhost mode, >>> and >>> AFAICT there isn't any way to get it to rewrite to the cache_peer name. >>> >>> Is this correct? If so, I might submit a patch to allow rewriting of the >>> Host header on a per-peer basis... >>> >>> Cheers, >>> >>> >>> -- >>> Mark Nottingham [EMAIL PROTECTED] >>> >>> >>> > > -- > Mark Nottingham [EMAIL PROTECTED] > > >
